My server rebooted for an unknown reason..

[J]

On Tue, May 25, 2010 at 13:09, Maxime Alarie <malarie at processia.com> wrote:
> Thanks Jeff,
>
>
> The "server" is a workstation,  its not a rackmount.  I checked the logs, and like you said, it might be a hardware issue, because no  shutdown messages were anounced in messages and syslog..  Only startup messages after some normnal operations.
>
> I will try to investigate the problem.
>
> Thanks much
>
> Maxime

Have fun ;-)  Given what you've said, sounds more like hardware than
OS (though an OS issue is not entirely out of the question).

One thing you can do that won't cost you anything is open the box and
reseat EVERYTHING.  I/O cables, power connectors, RAM sticks,
processors, etc.  Chip Creep is certainly a possibility and could be a
potential cause for issues like this, and the only thing it costs you
is the time to reseat stuff.

After that, well... who knows.  on a desktop, the BIOS MAY save some
previous error messages.

As for terminal programs...

powertop will give you an idea of how much power you're consuming...
if it's too high (assuming you can get ACPI power stats) you could be
seeing an overheating or failing power supply.

I'd say that if RAM is going bad, you may start to see kernel page
faults or other such oopses and warnings in the logs, though that's
not a guarantee.

hddtemp will show the temps of a given hard disk...

Also, what Clifford says about IPMI and smartmontools

Unfortunately, if the problem is HW related, it's usually pretty hard
to track down and the root cause can be non-obvious.

Another thing to consider... is the system on a UPS?  and could there
have been a power flicker over the weekend that would have been enough
to cause the system to lose power and restart?

I don't know about your part of the world, but I get that from time to
time when strong storm systems move through, and an abrupt loss of
power will also cause this kind of thing to happen.